Hello Sthu,
Thank You for Your time and extended answer, Joe:The general rule of thumb is to automatically update only workstations and test servers, never production servers. There is no way that automation can know whether it is wise to apply a particular update immediately, to wait a while, or not do it at all in cases of serious error. While errors (in Stable, at least) are extremely rare, the cost to a production server could be enormous.So, in case of workstations, we have no options besides these two: either risk or manual update - these are the options?
For production servers I have a small script** that alerts me via e-mail when there are new updates to install. I run this script nightly on at least one production server per type server I have. In the weeken I run it on alle servers just to make sure I do not forget to update a server.
On the first production server of a certain type I run the script manualy to see which prompts, if any, there are. If I see no strange things I can fire the automated update script on the other servers. HOW you do that is different per situation. I have a small enough number of servers per type that I can simply have a small script on my Windows workstation that does a remote launch using the plink command (from the PuTTY suite) per server.
** My UpgradeReport.sh script ----------<quote>----------------------------- #!/bin/bash MAILREC="emailaddress1@tio.nl emailaddress2@tio.nl" SUBJECT="Upgrade report voor $HOSTNAME" TMPFILE=/tmp/upgradereport.tmp # Stap 1: refresh repositories... apt-get update >/dev/null 2>&1 # Stap 2: show upgrades apt-get --dry-run upgrade | grep Inst > $TMPFILE if egrep -q ^Inst $TMPFILE then mail -s "$SUBJECT" $MAILREC < $TMPFILE fi rm $TMPFILE ----------<quote>----------------------------- Bonno